Elsevier

Hormones and Behavior

Volume 61, Issue 1, January 2012, Pages 31-36
Hormones and Behavior

Baby cries and nurturance affect testosterone in men

https://doi.org/10.1016/j.yhbeh.2011.09.012Get rights and content

Abstract

Testosterone (T) is generally theorized within a trade-off framework that contrasts parenting and low T with competitive challenges and high T. Paradoxically, baby cues increase T, prompting questions of whether T or its behavioral expression has been mischaracterized. We tested 55 men using a novel interactive infant doll paradigm, and results supported our hypotheses: We showed for the first time that baby cries do decrease T in men, but only when coupled with nurturant responses. In contrast, baby cries uncoupled from nurturant responses increased T. These findings highlight the need to partition infant cues and interactions into nurturant versus competitive-related contexts to more accurately conceptualize T, as per the Steroid/Peptide Theory of Social Bonds. This experiment also supports the utility of this paradigm for studying effects of infant interactions on hormonal responses, which may provide critical insights into ameliorating the darker sides of caregiving (e.g. anger, frustration, violence) and enhancing the positive sides (e.g. intimacy, nurturance, reward).

Highlights

► A reactive baby doll paradigm is effective at manipulating testosterone in men. ► Infant cries unaccompanied by nurturance increase testosterone. ► Infant cries accompanied by effective nurturance decrease testosterone. ► Baby cries elicit different testosterone responses depending on contextual factors. ► Parental behaviors can be partitioned into nurturant and competitive behaviors.

Introduction

Testosterone's association with social behaviors is generally understood as a trade-off between high T and competitive challenges versus low T and parenting. Evidence from a variety of species supports this distinction (Archer, 2006, Hirschenhauser and Oliveira, 2006, Ketterson and Nolan, 1992, McGlothlin et al., 2010, van Anders and Watson, 2006), which is often framed via the ‘Challenge Hypothesis’ (Wingfield et al., 1990). This trade-off is apparent over the lifespan, and parturition is associated with transient decreases in T for mammalian fathers (Nunes et al., 2000, Wynne-Edwards, 2001). Cross-sectional data from humans also suggest that mothers and fathers have lower T than their nonparent counterparts (Gray and Campbell, 2009, Kuzawa et al., 2009, Kuzawa et al., 2010) though only when they are involved in caregiving (Muller et al., 2009). Relatedly, two of the clearest predictions for T and social behaviors from the Challenge Hypothesis are that T should inhibit parenting, and infant cues should decrease T; this clarity is belied by contradictory data, as evidence shows that many infant cues and contexts actually increase T.

Studies on hormonal responses to baby cues in men find that baby cries increase T (Fleming et al., 2002, Storey et al., 2000), rather than decrease T in accordance with the low T/parenting link. Moreover, acute doses of exogenous T enhance neural responsivity to baby cries in women (Bos et al., 2010), rather than the suppressive effects that theory would predict. Yet, male trait T shows the expected associations such that lower baseline T is related to higher paternal empathy (Fleming et al., 2002). Similarly showing expected effects, experimental studies demonstrate that baby odor decreases T in marmoset fathers (Ziegler et al., 2011), more time with infants is associated with lower T in fathers (Storey et al., 2011), and T administration decreases psychological measures critical to empathy (Hermans et al., 2006). Why, then, does T responsivity to baby cues in humans diverge from longstanding understandings of T ?

One difficulty with interpreting research on T responses to baby cues is that the ‘parental/infant’ category is assumed to be phenomenologically whole. That is, all behavioral contexts related to infants or parenting are subsumed into one low-T category, regardless of their motivation or function. Activities as varied as breastfeeding, infant protection or defense, warm hugs, and hearing cries are categorized under the infant/parent umbrella. Yet infant defense clearly has a very different motivation than hugs, i.e. responding to a threat versus exhibiting nurturance, and a very different form in terms of observable behavior. Indeed, infant defense may be more akin to a challenge situation than one involving nurturance. According to the Steroid/Peptide Theory of Social Bonds, only those infant/parent contexts that involve nurturance will decrease T; those that involve competitions (real or imagined) will increase T (van Anders et al., 2011). Infant defense is typically studied in non-humans, and evidence from cichlid fish does show that it increases T (Desjardins et al., 2008). Might infant cries increase T for the same reason?

Baby cries might cue various behavioral responses in caregivers, including hugging or feeding. But in a situation where no nurturant response is possible, baby cries might actually cue danger or threat, and thus be more akin to a high T challenge situation. A hallmark of behavioral neuroendocrinology is that hormone–behavior associations are situated, i.e. that associations between hormones and behaviors are only visible given appropriate social contexts. Accordingly, both the stimuli and the response parameters (context and motivation) might be critical to understanding how baby cues modify T.

What is the behavioral role of T? Understanding why some infant cues increase T in men while others decrease it is critical to developing a more nuanced and comprehensive account of the function of T. Moreover, developing a paradigm to experimentally assess effects of baby interactions on adult physiology could be a novel tool for assessing parental behaviors and also useful for pinpointing where interventions might be usefully employed to aid caregivers in dealing with challenging baby interactions. The first major goal was to determine the extent to which baby cues and response context modulate T reactivity. We tested three hypotheses concerning effects of baby cues on adult T using a responsive baby doll: that (1) baby cries alone, with no opportunity for nurturant response, would increase T; (2) baby cries accompanied by effective nurturant responses that reduce cries would decrease T; and (3) baby cries accompanied by ineffective nurturant responses would lead to intermediate effects, and possibly no change in T. Here, the nurturant responses coupled with their ineffectiveness at abating the baby cries might cancel each other out. Alternatively, T might decrease given the nurturant responses, even though they were ineffective; or T might increase given the unabated cries. We focused on men because past research has only so far demonstrated these unexpected effects in men. A second major goal of this study was to assess a newly designed method for experimental tests of effects of baby cues and interactions on adult hormones.

Section snippets

Participants

There were 55 men in this study (mean age = 21.64 years, SD = 6.02), recruited through the university participant pool and the community via posters and advertisements. There were four fathers, each of whom had one, two, or three children between 4 years old and teen-aged (with one participant not indicating the child's age). Thirty-eight men identified having younger siblings, reporting one (n = 20), two (n = 14), three (n = 1), four (n = 1), or five (n = 1) younger siblings. Participants self-identified

Background Questionnaire

This addressed demographic factors, potential hormonal confounds, and experience with children.

Affect and Arousal Scale

We adapted a scale (Heiman, 1980) commonly used to measure affect and arousal before and after exposure to stimuli by adding four items (sad, happy, stressed, and relaxed) and removing 14 others related to sexuality. We have adapted this scale successfully in the past (Goldey and van Anders, 2011). The modified scale included 24 items that resolved into five subscales after Principal Components Factor

Confounds and covariates

Potential confounds in analyses with T include age, time of testing, date, nicotine use, exercise, and BMI (Gray et al., 2007, van Anders and Goldey, 2010). In addition, prior experience with infants might also affect T responses to infants (Fleming et al., 2002), and we recorded parental status, having or playing with younger siblings, babysitting in general or infants specifically, and babysitting siblings. Only exercise, date of testing, and sibling babysitting were meaningful covariates

Discussion

In this experiment, we examined effects of baby cues, context for response, and outcome on T in men. To do so, we developed a novel paradigm for examining effects of infant interactions on adult physiology using a realistic baby doll that responds both to participant behavior and internally programmed cues. Our results supported our three hypotheses as (1) baby cries alone significantly increased T in men when no nurturant response was possible; (2) baby cries comforted by nurturant responses

Acknowledgments

We would like to thank Catherine Marler and Eduardo Fernandez-Duque for helpful discussion, and the following for insightful comments on an earlier draft of this manuscript: Nicholas Caverly, Sara Chadwick, Patty Kuo, and especially Justin Carré, Katherine Goldey, Lauren Hipp, and Lane Ritchie. We would also like to thank Samantha Greenberg for baby doll troubleshooting, to Amanda Posh for characterizing the baby cry schedule in (excruciating) detail, and to our research assistants for being

References (33)

Cited by (135)

View all citing articles on Scopus
View full text